↓67 |
Our current knowledge of Arabidopsis mitochondrial promoters is largely based on computational predictions performed on the mitochondrial genome sequence of this plant. While a set of 29 hypothetical mitochondrial promoters has emerged from searching the Arabidopsis mtDNA for motifs known to drive transcription initiation in other dicotyledonous plants (Brennicke, et al., 1999;Dombrowski, et al., 1998), experimental evidence is limited to a single promoter in Arabidopsis mitochondria (Giese, et al., 1996). Here, the architecture and distribution of promoters in the upstream regions of twelve mitochondrial genes in Arabidopsis have been analyzed through mapping transcription start sites by a 5’-RACE technique that has previously been employed successfully for the analysis of bacterial and plastidial transcript 5’ termini (Argaman, et al., 2001;Bensing, et al., 1996;Miyagi, et al., 1998;Vogel, et al., 2003).
↓68 |
The method proved a valid tool for mitochondrial 5’ end detection, and for 19 out of 30 mapped promoters clearly discriminated between primary and processed transcript ends. However, for selected 5’ termini that were later identified through analysis of in vitro-capped transcripts to result from transcription initiation, 5’-RACE results did not support transcription initiation at the respective promoters. One example is Prrn18-69, for which the corresponding 5’-RACE product was significantly more abundant after amplification from transcripts not treated with TAP. Most likely, the misleading PCR result is due to a highly abundant primary transcript derived from the upstream promoter Prrn18-156, which is favoured in the 5’-RACE from TAP-treated transcripts, thereby outcompeting the 5’-RACE product generated from the much less abundant downstream 5’ terminus (compare rrn18 5’-RACE products in Figure 8 and in vitro-capped transcripts in Figure 12).
For a group of primary transcripts of the atp1, atp6-1, atp6-2, atp8 and atp9 genes mapping to CGTATATAA or CATAAGAGA motifs, 5’-RACE failed to distinguish primary from 5’-processed transcripts, owing to both types of transcripts mapping to the same initiating nucleotide. The processed transcripts could be derived from a modification of primary transcripts by a phosphatase or pyrophosphatase, or from endonucleolytic cleavage of transcripts initiated at upstream promoters.
Various other mechanisms of transcript 5’ end processing are implied from the analysis of mitochondrial RNA 5’ termini. Primary transcripts originating from Pcox2-210 and
Pcox2-481 appear to be trimmed by nucleolytic activities at several consecutive sites immediately downstream of the primary 5’ end, yielding processed transcripts of slightly different lengths. It is possible, however, that the shortened transcripts are not in vivo products but artefacts due to instability of specific transcripts during the experimental procedure. Several processed 5’ ends such as those mapping to positions -83 upstream of the atp9 coding sequence and -678 in the cox2 5’ region (compare Figure 31) are likely to have been generated through endonucleolytic removal of an RNA fragment from a newly initiated transcript. Endonucleases may be expected to function in mitochondrial transcript maturation in Arabidopsis, as all primary rRNA or tRNA 5’ ends defined in this study map to positions upstream of the 5’ terminus of the respective mature RNA. While 5’-terminal processing of tRNAs in plant mitochondria has been established to be carried out by RNase P or its homologue, enzymes involved in 5’-end maturation of rRNAs or mRNAs are as yet unidentified (Binder and Brennicke, 2003;Marchfelder, et al., 1996).
↓69 |
Nearly 50% of the transcriptional starts identified in this study are located within sequences that conform to the nonanucleotide consensus CRTAAGAGA previously suggested for dicot mitochondrial promoters (Binder, et al., 1996) or the motif CGTATATAA defined in this work (Table 9). About as many promoters deviate to a varying extent in the sequence surrounding the transcriptional start and in place of a CRTA core display ATTA or RGTA tetranucleotides that have emerged from this study as frequent promoter elements. 20 out of 30 promoters support initiation at an adenine nucleotide, which complements previous reports on transcription initiation mostly at guanine nucleotides in dicot mitochondria (Fey and Marechal-Drouard, 1999). While purines appear obligatory at positions +1 and +2 with respect to the transcriptional start, they are moderately frequent at positions +3 to +8. Preceding the promoter core is usually a sequence rich in A and T nucleotides, which has been described for several mitochondrial genes in dicots as well as in maize to be important for unimpeded promoter function in vitro(Dombrowski, et al., 1999;Rapp, et al., 1993). Moreover, the all-A/T sequence TATATA is seen as an element not only of the CGTATATAA motif but also of deviating promoters such as Prrn18-353 and Patp6-1-916/913 (Table 9), emphasizing the predisposition of A/T-rich nucleotide sequences to function as promoters. In line with this idea is the observation that of the two promoters Patp8-710 and PtrnM-574/573 showing neither a recognizable core motif nor the TATATA element, the latter is highly rich in A-T base pairs. The ability of mitochondrial sequences composed entirely of A and T nucleotides to support transcription initiation has been pointed out by (Lupold, et al., 1999), based on their own observation of an all-A/T promoter of the maize cox2 gene and an earlier report on yeast petite mutants presumably initiating mitochondrial RNA-synthesis at all-A/T sequences (Fangman and Dujon, 1984).
The detection of transcription initiation sites in the Arabidopsis mtDNA was strongly biased towards the identification of promoters displaying a CRTAAGAGA or CGTATATAA motif, as 9 out of 12 genes were selected for the analysis of transcriptional starts owing to the prior observation of at least one of these motifs in their 5’ regions. Any of these motifs were confirmed to be part of an active promoter. Yet, since 50% of the identified promoters were found to deviate from these motifs, considerably more than half of all mitochondrial promoters in Arabidopsis may be expected to show divergent sequences. This underlines the necessity to experimentally define mitochondrial transcription initiation sites and the limited possibility of predicting these sites, based on conserved promoter motifs.
Defined promoter elements such as particular core sequences or the TATATA motif appear to be distributed randomly between different promoters. This precludes Arabidopsis mitochondrial promoters from being classified into distinct groups that could be related to different mitochondrial RNA polymerases, based on merely a comparison of promoter sequences. In Arabidopsis, the nucleus-encoded phage-type RNA polymerase RpoTmp is targeted not only to mitochondria but also to plastids (Hedtke, et al., 2000). This enzyme could thus be expected to recognize promoters of similar structure in both organelles. Many of the mitochondrial promoters characterized in the present study contain a CRTA core, which resembles the YRTA motif displayed by a subset of plastid promoters that are most likely used by phage-type RNA polymerases (Hess and Börner, 1999;Liere, et al., 2004). Yet, the relatively small number of plastid promoters of this type studied thus far and the variability of mitochondrial promoter structures render it difficult to tentatively assign a distinct subset of mitochondrial promoters to RpoTmp. Besides, promoter recognition by this RNA polymerase may be mediated by different, as yet unidentified cofactors in plastids and in mitochondria, and may therefore depend on different promoter sequences in the two organelles.
↓70 |
To study the relevance of distinct sequence elements or nucleotide positions for promoter activity, transcription from mitochondrial promoters and mutagenized promoter variants might be analyzed using an in vitro transcription system as described previously for maize and pea mitochondrial promoters (Dombrowski, et al., 1999;Rapp, et al., 1993). Alternatively, comparative promoter analyses might be carried out in isolated organelles. Techniques for the introduction of DNA into isolated mitochondria from maize, sorghum and cauliflower have been developed, and the transcription of introduced genes has been shown (Farre and Araya, 2001;Staudinger, et al., 2005;Staudinger and Kempken, 2003); F. Kempken, Universität Kiel, personal communication).
Frequent duplications and rearrangements of plant mitochondrial genomes during evolution (Palmer, 1990) appear to have been important mechanisms of establishing promoter sequences throughout the mtDNA that are recognized by the nucleus-encoded transcription apparatus. Extended sequences showing high similarity to the rrn18 5’ region and comprising both Prrn18-69 and Prrn18-156 are seen upstream of the orf275-orf149-nad5c-nad4L-orf25 and orf153b-orf118-orf114-nad3-rps12-orf117-ccb203 gene clusters. Although it was not tested whether transcription is initiated within the duplicated rrn18-like regions, it is possible that they direct transcription of the two clusters. A duplication of promoter regions is most apparent for the atp1, atp6, atp8, atp9 and rrn26 genes, which possess homologous 5’ sequence stretches that comprise one or two promoters (compare promoter sequences in Table 9 and promoter distribution in Figure 31).
The idea has been put forward that in plant mitochondria, different promoter types, which could possibly be activated by distinct transcriptional cofactors, might control the transcription of particular genes or groups of genes (Mackenzie and McIntosh, 1999). However, promoter type distribution in Arabidopsis mitochondria appears to be at random, in part owing to the high frequency of mtDNA recombination events, and is thus not in support of such a mode of transcriptional control. For example, no differences are apparent between promoters of protein- or RNA-coding genes, or between genes encoding ribosomal proteins or components of the respiratory complexes.
↓71 |
For 9 out of 12 genes, multiple transcription initiation sites were detected (Figure 31). Six genes were found to be transcribed from three or even four promoters. Interestingly, in a series of initiation sites of one gene, promoters containing the conserved nonanucleotide motifs CGTATATAA or CRTAAGAGA were usually positioned downstream of promoters with deviating or apparently no motifs. It may be speculated that transcription is initiated at additional sites even further upstream of the defined promoters, but that it has been impossible to detect these as the experimental strategy applied in this study does probably not cover regions exceeding 2 kbp. Identifying additional upstream sites would require an extensive primer walking strategy.
While in dicots, multiple promoters have so far only been reported for the pea cox2 gene (Kuhn and Binder, 2002), transcription initiation at multiple sites has been described for several mitochondrial genes in maize, rice and sorghum (Lupold, et al., 1999;Mulligan, et al., 1988;Nakazono, et al., 1996;Nakazono, et al., 1996;Yan and Pring, 1997). It is unlikely that frequent promoter multiplicity, as revealed here for the first time for a dicotyledonous plant, would be restricted to Arabidopsis. 5’-RACE analyses, which easily trace even small amounts of primary transcripts, might uncover multiple promoters also in the mitochondria of other dicots.
↓72 |
Within the regions investigated for transcription start sites, additional promoters not displayed in Table 9 may be active from which minor primary transcripts derive. Transcripts mapping to the possible transcription initiation sites Pcox2-231 (AGTATTATAA, see Figure 10) and Prps3-1149 (AGTAGATAG) did not give rise to distinct bands in the 5’-RACE, but were revealed as by-products of cloning and extensive sequencing of more abundant 5’-RACE products resulting from nearby transcriptional starts. These possible promoters are not listed in Table 9 because not enough data has been accumulated to substantiate transcription initiation at these sites. The cloning of transcript 5’ ends not producing signals above the limit of detection in in vitro capping (Pcox2-231) or 5’-RACE experiments implies that a multitude of minor transcription start sites within mitochondrial DNA sequences may exist.
The role of mitochondrial promoter multiplicity in plant mitochondria is hitherto unknown. While variation in mitochondrial promoter usage has been described between Zea perennis plants possessing different alleles of a nuclear gene (Newton, et al., 1995), there has been no evidence to date that in a single plant species with a distinct nuclear background, mitochondrial genes would be differentially transcribed. The present study shows mitochondrial transcription to initiate at identical sites in leaves and in flowers of Arabidopsis plants, indicating that tissue-specific regulation of mitochondrial genes on the level of promoter selection is of only minor or no importance in Arabidopsis. This is in line with previous conclusions that in mitochondrial gene expression, regulatory mechanisms are aimed predominantly at posttranscriptional steps (Giegé, et al., 2000). Different promoters of a distinct gene may, however, vary in their level of activity between different tissues and possibly also between different developmental stages. Quantitative analyses of primary transcripts through primer extension experiments or quantitative real-time PCR would contribute to elucidating the activity and function of individual promoters controlling one gene.
It has been suggested that multiple promoters might enable regulatory mechanisms such as differential promoter usage, producing different 5’-untranslated regions which possibly influence translational yield (Lupold, et al., 1999). Such mechanisms would be contrasted by analyses that detected a similar distribution of cox3 transcripts with different 5’-untranslated sequences in both polysomal fractions and total RNA prepared from maize mitochondria, implying that mitochondrial ribosomes non-preferentially associate with differently initiated transcripts (Yang and Mulligan, 1993). From the present data on primary transcript 5’ termini in Arabidopsis mitochondria, a point of view is favoured that considers multiple promoters the result of a relaxed promoter specificity of the mitochondrial transcription machinery (Lupold, et al., 1999). Further experiments employing a defined Arabidopsis in vitro transcription system will be required to clarify whether specific promoter types are recognized by different RNA polymerase complexes containing particular transcriptional cofactors or different core enzymes.
↓73 |
A number of transcription initiation sites were detected in the present study that give rise to antisense transcripts to known genes or drive the expression of regions lacking identified or hypothetical ORFs, and are active in both leaves and flowers of Arabidopsis. Although it is possible that non-coding or antisense transcripts have a function in Arabidopsis mitochondria, experimental data in support of such a concept are lacking to date. It may rather be that these RNAs are the result of a relaxed transcriptional control and of non-stringent sequence requirements of plant mitochondrial promoters. In line with this, Arabidopsis mitochondrial promoters are highly diverse in architecture (compare IV.1.2), and multiple promoters driving the transcription of the same gene are apparently indifferently active (compare IV.1.3). It is possible that plant mitochondria, which undergo recurrent partial duplications and structural rearrangements of their genomes (see I.2), have accumulated transcription start sites at a fairly high frequency on the mtDNA, resulting in the synthesis of both coding and non-coding transcripts. Transcription of large non-coding regions has been described earlier in maize mitochondria (Finnegan and Brown, 1990). Although the authors did not determine whether these RNA products were due to unattenuated transcription of coding sequences or resulting from independent transcription initiation events, it is conceivable that the maize mitochondrial genome is like the Arabidopsis mtDNA transcribed from a multitude of promoters not confined to upstream regions of functional ORFs. Synthesis of transcripts from intergenic regions and of antisense RNAs has lately been reported in a variety of organisms (Johnson, et al., 2005) and is evidently not restricted to the mitochondrial compartment. Hence, as opposed to an earlier suggestion by (Lupold, et al., 1999), it may not necessarily be in response to structural peculiarities of the mtDNA that promoter multiplicity is maintained in plant mitochondria.
Support for non-functional and possibly disadvantageous transcripts being made in plant mitochondria, the synthesis of which is compensated by RNA degradation mechanisms, has recently been provided by (Holec, et al., 2005) who investigated mitochondrial RNA degradation substrates. In transgenic Arabidopsis down-regulated for polynucleotide phosphorylase (PNPase), RNAs targeted by polyadenylation for degradation by PNPase were found to include RNAs transcribed from regions that lack known functional genes (Holec, et al., 2005). Among these were transcripts of chimeric ORFs created by intragenic recombination events in the mitochondrial genome (compare I.2) as well as antisense transcripts to functional genes, which unless degraded might impede mitochondrial function. Selected RNAs were demonstrated by Northern hybridization analyses to significantly accumulate in the PNPase mutant but not in wild-type plants, and were detected in the wild type only by using more sensitive PCR techniques (Holec, et al., 2005). Plants with reduced PNPase levels were previously described to exhibit developmental defects (Perrin, et al., 2004), which may in part be due to accumulating deleterious transcripts (Holec, et al., 2005).
The seemingly non-stringent transcriptional control of plant mitochondrial gene expression thus appears to be contrasted and is possibly made up for by posttranscriptional processes. This concept may be valid for other genetic systems as similar mechanisms have been inferred from a study on transcription initiation sites in E. coli. Promoter-like sequences located within or downstream of ORFs on the E. coli chromosome were shown to activate transcription but to give rise to only minor RNA products, which might be a consequence of either imperfect promoter architecture or poor transcript stability (Kawano, et al., 2005).
↓74 |
Mitochondrial transcription in yeast depends on the cofactor mtTFB in addition to the phage-type RNA polymerase Rpo41 (Lisowsky and Michaelis, 1988;Masters, et al., 1987;Schinkel, et al., 1988;Winkley, et al., 1985). An RNA polymerase of the phage type has been identified as the catalytic subunit of the transcription machinery in human mitochondria (Tiranti, et al., 1997), followed by the discovery of two human genes encoding mitochondrial mtTFB-like factors that induce promoter-specific transcription in vitro(Falkenberg, et al., 2002;McCulloch, et al., 2002). According to secondary and tertiary structure analyses, mtTFB is a member of the group of S-adenosyl-l-methionine-dependent rRNA methyltransferases (Falkenberg, et al., 2002;McCulloch, et al., 2002;Schubot, et al., 2001). No such factors have so far been described in plants although plant mitochondria similarly rely on phage-type RNA polymerases to transcribe their genomes and are likely to require transcriptional cofactors (Hess and Börner, 1999), see I.3.3). The present study aimed at the identification of sequences encoding candidate auxiliary factors of mitochondrial RNA polymerases in Arabidopsis.
BLAST searches detected three nuclear genes in this plant encoding putative rRNA dimethylases with limited similarity to yeast and human mtTFB. Among these was PFC1, which has been reported earlier to encode a plastidial 16S rRNA dimethylase that is homologous to the yeast nucleolar 18S rRNA dimethylase Dim1 (Tokuhisa, et al., 1998). GFP import experiments have confirmed PFC1 to possess an N-terminal transit peptide mediating protein import into plastids of isolated tobacco protoplasts (B. Kuhla, HU Berlin, personal communication). Two other mtTFB-like genes corresponding to loci At5g66360 and At2g47420 were tentatively designated MetA and MetB respectively. Analyses of subcellular targeting demonstrated the derived MetA protein to possess an N-terminal transit peptide mediating translocation into mitochondria, thus supporting a mtTFB-like localization of MetA, whereas MetB may localize to the cytoplasm or nucleus.
Recombinant MetA and MetB expressed in and purified from E. coli were found to bind DNA but displayed no specificity for mtDNA fragments containing mitochondrial promoters. Not only MetA but also the presumably non-mitochondrial MetB behaved identically to human mtTFB1 in DNA binding assays (compare (McCulloch, et al., 2002). It thus appears that non-specific DNA binding, which has also been reported for yeast mtTFB (Riemen and Michaelis, 1993), is a general attribute of rRNA dimethylases and rRNA-dimethylase-like proteins that is not related to the function of some members of the methyltransferase group as mitochondrial transcription factors. This idea is in line with a report on a point mutation in
h-mtTFB1 that eliminated DNA binding but did not affect transcriptional activation by
h-mtTFB1 in vitro(McCulloch and Shadel, 2003). The authors accordingly proposed a model in which h-mtTFB1 or h-mtTFB2 does not itself bind to the promoter but bridges an interaction of h-mtTFA bound to distal promoter elements with the core RNA polymerase at the transcription initiation site (see I.3.3.2 and Figure 6).
↓75 |
In order to resolve the relationships of MetA, MetB and Pfc1 to yeast and animal mtTFBs and to characterized rRNA dimethylases, a phylogenetic analysis was performed, which also included rRNA dimethylase-like sequences available from plants other than Arabidopsis. In line with targeting analyses (see III.2.2, Figure 16, and Annex B), Arabidopsis MetB and its supposed nuclear or cytosolic plant orthologues were calculated to be most closely related to a group of rRNA dimethylases containing the fungal nucleolar rRNA dimethylase Dim1 as well as an additional, possibly nuclear or cytoplasmic methyltransferase from humans.A group of predicted mitochondrial rRNA dimethylases of plants, which include Arabidopsis MetA, appear to have separated from the line leading to the MetB/Dim1 cluster significantly more recently than fungal mtTFB and animal mtTFB1 and mtTFB2. Consistent with the phylogenetic relation between methyltransferase-like sequences (Figure 17), it is possible that the evolutionary process leading to the recruitment of an rRNA dimethylase as auxiliary factor of the mitochondrial transcription machinery in fungi and animals took place after the separation of the common ancestor of these groups from the green algae and plant lineage (for eukaryotic phylogeny, see (Burger, et al., 2003). Thus, plant mitochondrial rRNA dimethylase-like proteins need not necessarily represent transcriptional cofactors. In vitro transcription studies as conducted for the human mitochondrial transcription machinery (Falkenberg, et al., 2002) and analyses of transgenic Arabidopsis lines impaired in MetA function might provide the means to clarify whether Arabidopsis MetA acts as mitochondrial transcription factor.
The position of Arabidopsis and poplar Pfc1 in the phylogram (Figure 17) is similarly inconclusive with respect to a possible additional role of these proteins as cofactors of phage-type RNA polymerases in plastids. Pfc1 appears to have separated from the line leading to the MetB/Dim1 group much earlier than MetA but is not more closely related to fungal mtTFB or animal mtTFB1 and mtTFB2 than MetA. However, in the light of the extreme divergence of fungal and animal mtTFBs, evolutionary distance to these proteins does not argue against a role of Pfc1 in plastidial transcription.
Surprisingly, fungal and animal mtTFBs do not form a single cluster in the phylogram (Figure 17). It moreover appears that mtTFB1 and mtTFB2 sequences, each forming a separate cluster containing exclusively animal proteins, are less closely related to each other than any of these groups to a cluster containing fungal nucleolar rRNA dimethylases and their putative animal orthologues. Compared to mtTFB1s, the mtTFB2 group is moderately more closely related to the highly diverse fungal mtTFBs. It thus appears that in contrast to a previous suggestion (Falkenberg, et al., 2002), mtTFB1 and mtTFB2 are not the result of a gene duplication event early in metazoan evolution and instead have separated at a prior point in time. It has been reported that human mtTFB1 is an order of magnitude less active as transcription factor than human mtTFB2 (Falkenberg, et al., 2002); the former is able to methylate rRNA at a conserved methylation site (Seidel-Rogol, et al., 2003). These observations motivated the proposition that despite the similarity of both proteins to rRNA dimethylases and their ability to act as transcription factor in vitro, the methyltransferase and cofactor roles could be distributed between mtTFB1 and mtTFB2 in human mitochondria (McCulloch and Shadel, 2003). Support for this idea of biochemically non-redundant functions of mtTFB1 and mtTFB2 is gained from two studies on the roles of mtTFB1 and mtTFB2 in Drosophila. While cells with an RNAi-mediated reduction in mtTFB2 levels displayed defects in mitochondrial transcription and replication that could not be compensated by mtTFB1 (Matsushima, et al., 2004), RNAi knock-down of mtTFB1 reduced mitochondrial protein synthesis but had no effect on the abundance of specific mitochondrial transcripts (Matsushima, et al., 2005). The proposed roles of animal mtTFB2 as mitochondrial transcription factor and of mtTFB1 as translational modulator (Matsushima, et al., 2005) are in accordance with the phylogenetic analysis presented here.
↓76 |
It has been suggested earlier that the methyltransferase properties of mtTFB and its ability to bind RNA may be important for transcription factor activity (Schubot, et al., 2001). However, it has been demonstrated for human mtTFB1 that the in vitro function of the protein as transcription factor is independent of its rRNA dimethylase activity (McCulloch and Shadel, 2003). This emphasizes that the evolutionary process turning an rRNA dimethylase into a cofactor of a phage-type RNA polymerase may have been promoted by structural features of the methyltransferase but may not have depended on methyltransferase activity.
Considering the poor amino acid sequence conservation among fungal and animal mtTFBs (see III.2.1 and Annex A), the possibility must be considered that additional mtTFB-like sequences exist in plant genomes that encode organellar rRNA dimethylases or/and transcription factors, but which the BLAST searches conducted here have failed to detect.
The homology of nucleus-encoded plastidial and mitochondrial RNA polymerases in plants and the presence of yet another phage-type RNA polymerase in both mitochondria and plastids in dicotyledonous plants (Chang, et al., 1999;Hedtke, et al., 1997;Hedtke, et al., 2000;Hedtke, et al., 2002;Ikeda and Gray, 1999) raise the question whether the two organelles harbour similar transcriptional cofactors interacting with these core enzymes. As yet, the only characterized gene encoding a plastidial protein with similarity to known mitochondrial transcription factors in eukaryotes is the mtTFB-like rRNA dimethylase gene PFC1.
↓77 |
A transgenic Arabidopsis line carrying a T-DNA insert in PFC1 has been reported to lack 16S rRNA methylation in the plastid at two adenosines established as conserved modification sites of the ribosome small subunit rRNA (Lafontaine, et al., 1994;Tokuhisa, et al., 1998). The characterization of this mutant had been carried out before the resolution of the three-dimensional structure of yeast mtTFB and the identification of the gene encoding human mtTFB1 uncovered these organellar transcription factors to be related to rRNA dimethylases (McCulloch, et al., 2002;Schubot, et al., 2001). Hence, the pfc1 mutant line was not examined for possible alterations in transcripts synthesized in the plastid. Pfc1-deficient plants were of a phenotype identical to wild-type individuals at 22°C but exhibited chilling-induced chlorosis at 5°C due to impaired chloroplast development at low temperature and returned to normal chloroplast biogenesis in meristematic cells upon transfer to 22°C (Tokuhisa, et al., 1998). Pfc1 was thus concluded to be essential for chloroplast development at low temperature and to possibly compensate for a chilling-sensitive step in ribosome biogenesis. The protein may thus have a role in plastids that is equivalent to the suggested in vivo function of Drosophila mtTFB1 in mitochondria (see I.3.3.1 and IV.2.2). Further studies will be required to examine if in addition to being an rRNA dimethylase, Pfc1 is a cofactor of phage-type RNA polymerases in plastids, and if the described low-temperature defects in pfc1 are at least partially due to imperfect transcription in the photosynthetic organelle.Analyses of transcripts in pfc1 plastids are under way (M. Swiatecka, HU Berlin, personal communication).
No sequence encoding a putative plastidial Pfc1-like protein could be retrieved from the fully sequenced rice genome or from any other monocot species, indicating that 16S rRNA maturation and RpoT-driven transcription in monocot plastids might be independent of a methyltransferase-like protein. Alternatively, Pfc1-like sequences of monocots may have escaped the BLAST searches conducted here (compare IV.2.2).
The RNA polymerases and auxiliary factors involved in mtDNA transcription in yeast and mammals have been identified and subjected to thorough analyses of their contribution to the transcription process (Falkenberg, et al., 2002;Gaspari, et al., 2004;Matsunaga and Jaehning, 2004); for reviews summarizing earlier studies, see (Hess and Börner, 1999;Shadel and Clayton, 1993;Tracy and Stern, 1995). Reconstitution of mitochondrial transcription in vitro from individual recombinant components allowed defining the minimal composition of RNA polymerase complexes capable of mitochondrial promoter recognition and transcription initiation (Falkenberg, et al., 2002;Gaspari, et al., 2004;Matsunaga and Jaehning, 2004;Matsunaga, et al., 2004). The activity of the transcription apparatus present in plant mitochondria, on the other hand, has so far mostly been characterized by analyses of transcripts synthesized in vivo in the mitochondrion (Hess and Börner, 1999), and references therein), or through detailed studies of in vitro RNA synthesis by transcriptionally active mitochondrial extracts (Binder, et al., 1995;Hanic-Joyce and Gray, 1991;Rapp and Stern, 1992). Although a biochemical dissection of such extracts has been anticipated to lead to the identification of individual components of the transcription machinery, as was attained earlier for the yeast mitochondrial RNA polymerase and its transcriptional cofactor sc-mtTFB (Greenleaf, et al., 1986;Lisowsky and Michaelis, 1988;Schinkel, et al., 1987;Winkley, et al., 1985), direct evidence is as yet lacking for specific components to be involved in mitochondrial transcription in plants. Attempts to isolate the core RNA polymerase and accessory factors have been hampered presumably mostly by the low abundance of these components in the mitochondrion and by their seemingly loose association in RNA polymerase holoenzymes.
↓78 |
The identification of mitochondrial promoters (this study; (Kühn, et al., 2005) and of genes encoding RNA polymerases (Hedtke, et al., 1997;Hedtke, et al., 2000) and a potential mtTFB-like cofactor (this study) of the mitochondrial transcription machinery in Arabidopsis prompted efforts to reconstitute a defined homologous in vitro transcription system from individual recombinant components.
In vitro transcription studies examined the abilities of RpoTm and RpoTmp to transcribe DNA from mitochondrial promoters located on supercoiled or linearized plasmid templates, and moreover tested the mitochondrial mtTFB-like protein MetA for its potential to modulate the transcriptional performances of RpoTm and RpoTmp. Transcription of neither supercoiled nor linear DNA templates by RpoTm or RpoTmp appeared to be affected by the presence of MetA in the in vitro assay. This contrasts the functioning of yeast mtTFB which stimulates promoter-specific transcription of both linear and supercoiled DNA (Matsunaga and Jaehning, 2004;Winkley, et al., 1985). A role of MetA that would be equivalent to that of yeast or mammalian mtTFB in mitochondrial RNA synthesis is thus not supported by the in vitro transcription experiments described here (see IV.4). While RpoTm recognized a variety of promoters on supercoiled DNA, identical promoters did not yield run-off RNA products if provided on linearized templates (Table 12 and Figure 23, Figure 25, Figure 28 and Figure 30). Likewise, RpoTmp produced no specific RNAs from linear DNA. The synthesis of discrete transcripts from supercoiled plasmids by RpoTmp was confined to the promoters Patp6-1-916/913, Patp6-2-436 and Patp6-2-507 (Figure 28 and Figure 30). Table 12 lists promoter sequences that were included in the study and indicates their in vitro utilization.
In vitro transcription studies made use of the two bacterial ρ-independent terminator sequences hisa and thra(Barnes and Tuley, 1983;Gardner, 1982) in order to obtain RNA products of defined lengths resulting from promoter-specific initiation followed by termination at hisa or thra on circular templates (see Figure 23, Figure 25, and Figure 27 for template design). ρ-independent terminators resemble Class I terminators of the T7 phage in that they encode a G/C-rich sequence stretch followed by a run of U residues, with the former having the potential to form a stable stem loop (Farnham and Platt, 1981;Macdonald, et al., 1994;Macdonald, et al., 1993;von Hippel, 1998). A number of bacterial termination signals were shown accordingly to cause the T7 RNA polymerase to terminate (Christiansen, 1988;Jeng, et al., 1990;Jeng, et al., 1992;Steen, et al., 1986). The bacterial terminators hisa and thra seem to similarly induce termination of transcription by the T7-like enzymes RpoTm and RpoTmp in vitro although termination in vivo does apparently not involve terminator sequences in mitochondria (see I.3.2.3).
↓79 |
Previously described in vitro transcription studies routinely made use of primer extension analysis in order to characterize in vitro transcription product 5’ ends (Binder, et al., 1995;Liere and Maliga, 1999). However, occasional discrepancies have been observed between in vitro transcription and primer extension signals (Binder, et al., 1995). Therefore, precise 5’ end mapping was performed here on in vitro-synthesized RNAs employing the 5’-RACE technique used for the detection of in vivo transcription initiation sites in Arabidopsis mitochondria (see III.1.1). The method proved a valid tool for determining transcript
5’ termini of in vitro transcription products.
Table 12: Promoter recognition by RpoTm and RpoTmp in vitro.
Gene |
Promoter |
Sequence |
RpoTm |
RpoTmp |
rrn18 |
Prrn18-156 |
tagaataatacg tatatAatcagaa |
+ |
- |
atp6-1 |
Patp6-1-200 |
gccaataatacg tatatAagaagag |
+ |
- |
atp9 |
Patp9-295 |
ctggtgctctcg tatatAagagaag |
+ |
- |
cox2 |
Pcox2-210 |
atgttggtttcg tatatAagaagac |
+ |
- |
tRNA-fMet |
PtrnM-98 |
tttgaaatatcgtaagaGaagaagg |
+ |
- |
rrn26 |
Prrn26-893 |
ctatcaatttcataagaGaagaaag |
+ |
- |
atp9 |
Patp9-239 |
ctatcaatttcataagagaagacga |
- |
- |
atp6-1 |
Patp6-1-156 |
ctatcaatctcataagagaagaaat |
- |
- |
atp8 |
Patp8-157 |
ctatcaatctcataagaGaaGaaat |
(+)a |
- |
rrn18 |
Prrn18-69 |
agtggaattgaataagagaagaaag |
- |
- |
atp8 |
Patp8-228/226 |
cataccataaca tatatAgaatcga |
+ |
- |
atp6-1 |
Patp6-1-916/913 |
agccctttatat tatatAatAaagc |
+ |
+ |
cox2 |
Pcox2-481 |
atgaatattcattagatAataGatt |
+ |
- |
rps3 |
Prps3-1133 |
tagaaaaaattattagtaatacgta |
- |
- |
rps3 |
Prps3-1053 |
ttttttatttggtaggtaacatcgc |
- |
- |
atp9 |
Patp9-487 |
atgtcttattggtatgtGatacaag |
+ |
- |
atp9 |
Patp9-652 |
agaagattgaagtaaggagcaggtt |
- |
- |
cox2 |
P*cox2-345 b |
tctgtactgtagtaataGaagagtc |
+ |
- |
atp6-2 |
Patp6-2-436 |
tcttgaattaag tatataGaaaaga |
+ |
+ |
atp6-2 |
Patp6-2-507 |
gataaattaagtatagtaatAagaa |
+ |
+ |
tRNA-fMet |
PtrnM-574/573 |
ctaatttatataaaaaagaccggga |
- |
- |
Of the promoters tested, all sequences displaying a TATATA element were recognized by RpoTm in vitro. Similarly, all promoters with a CGTA core stimulated transcription initiation, although PtrnM-98, which is the only tested promoter possessing a CGTA motif but not a TATATA sequence, appeared to support initiation less efficiently. Of the promoters displaying a CATAAGAGA nonanucleotide motif, only Prrn26-893 gave rise to discernible though not abundant transcripts. Run-off transcription and 5’ end mapping of in vitro-synthesized RNAs together provided no evidence for transcripts initiated specifically at Patp6-1-156, Patp9-239or the similar Prrn18-69, and a minor stimulation of transcription initiation by Patp8-157 was revealed only by the sensitive 5’-RACE technique (compare Figure 28 and Figure 29), indicating that RpoTm does not significantly prefer these in vivo promoters over random start points in vitro. Since despite the high sequence similarity at these promoters to Prrn26-893, transcription is detectably initiated only at Prrn26-893, it may be assumed that already minor deviations from the Prrn26-893 sequence impede recognition by RpoTm, or that sequences beyond the 25 nucleotides displayed in Table 12 determine promoter strength in vitro. Interestingly, upstream sequences extending beyond the T7 RNA
polymerase promoter boundary at position -17 were recently shown to modulate promoter efficiency (Tang, et al., 2005). In vivo utilization of not only Patp6-1-156, Patp9-239 and Prrn18-69 but also Prrn26-893 might depend on the function of as yet unidentified transcriptional cofactors and/or extended sequence elements.
↓80 |
Earlier in vitro transcription studies using transcriptionally active extracts prepared from pea mitochondria described the faithful initiation at several promoters essentially conforming to the CRTAAGAGA consensus motif but rarely at promoters deviating from this sequence (Binder, et al., 1995;Kuhn and Binder, 2002). In the present study, in vitro utilization by RpoTm of promoters not possessing a CRTA sequence element but an ATTA or RGTA core was heterogeneous. While transcripts initiated at Patp6-1-916/913 and Patp9-487 had
5’ termini identical to those of transcripts made from these promoters in vivo, initiation at Pcox2-481 occurred at the correct nucleotide but in addition at a nearby downstream site not utilized in vivo. Transcript 5’ ends generated from Patp6-2-436 and Patp6-2-507 mapped to positions closely downstream of the in vivo initiating nucleotides. This may be either due to an altered or diversein vitro activity of RpoTm as compared to the in vivo initiation event, or caused by intrinsic problems of the 5’-RACE procedure at the ligation step. None of the rps3 promoters gave rise to specific transcripts; neither did Patp9-652 or the A/T-rich PtrnM-574/573 lacking apparent promoter sequence elements (compare IV.1.2.). The failure of both RpoTm and RpoTmp to initiate transcription at certain promoters may be due to the absence of factors activating transcription in vivo in the Arabidopsis in vitro transcription system. Such factors need not necessarily differ between dissimilar promoter sequences.
Minor transcripts that had not been detected in vivo were initiated by RpoTm in vitro at several positions of the atp9 upstream region. Besides, one major discrete but non-specific transcript was made from the cox2 upstream sequence. Promoter function of an mtDNA sequence in vitro but not in vivo has been observed previously in in vitro studies employing maize mitochondrial extracts as a source of transcription activity (Lupold, et al., 1999). Of several primary transcript 5’ ends generated in vitro from two different maize mtDNA regions harbouring the cox2 sequence, all except one were confirmed to also occur in vivo. The authors argued that rapid processing of this transcript in vivo might account for the discrepancy between in vitro transcription results and transcript 5’ ends seen in vivo, but did not exclude the possibility that the sequence does not serve as a promoter in the mitochondrion.
Earlier reports have put forward that not only RNA polymerase binding to promoter (and promoter-like) sequences but also escape from these sequences into elongation may be limiting to mitochondrial promoter function and transcription initiation in vivo(Lupold, et al., 1999;Matsunaga and Jaehning, 2004). According to a recent study of the yeast mitochondrial transcription machinery, sc-mtTFB may participate in modulating this transition. Yeast Rpo41 was shown to possess the intrinsic ability to correctly initiate transcription at mitochondrial promoters on supercoiled or pre-melted DNA templates in the absence of the accessory factor sc-mtTFB, whereas addition of the cofactor appeared to increase abortive relative to productive transcription from pre-melted templates (Matsunaga and Jaehning, 2004). Specific interaction with promoter sequence elements has moreover been attributed to the human and mouse mtTFA and mitochondrial RNA polymerase polypeptides rather than mtTFB (Gaspari, et al., 2004); compare I.3.3.2 and Figure 6). The methyltransferase-like transcriptional cofactors of animal and yeast mitochondria, which are obligatory for specific transcription initiation on linear DNA templates (Falkenberg, et al., 2002;McCulloch, et al., 2002;Winkley, et al., 1985;Xu and Clayton, 1992) owing to their contribution to DNA melting (Matsunaga and Jaehning, 2004), and which have previously been discussed as specificity factors involved in promoter recognition (Jang and Jaehning, 1991;Xu and Clayton, 1992), may thus alternatively be regarded as modulators of open promoter complex stability and escape into productive transcription from mitochondrial promoters. Correspondingly acting auxiliary factors may be part of plant mitochondrial transcription machineries, and their absence in the Arabidopsis in vitro transcription system may result in erratic initiation events.
↓81 |
In line with observations of an intrinsic promoter specificity of the yeast mitochondrial RNA polymerase Rpo41 is the accurate transcription initiation by RpoTm at diverse promoters of the Arabidopsis mtDNA, which is described here. RpoTm differs in its transcriptional performance from complex plant mitochondrial extracts in that it does not specifically initiate transcription at promoters located on linear DNA templates. This implies the participation of (an) auxiliary factor(s) in alterations of the DNA structure or in template melting and open promoter complex formation in in vitro transcription experiments using mitochondrial extracts (Binder, et al., 1995;Hanic-Joyce and Gray, 1991;Rapp and Stern, 1992), which according to the data presented here is obviated by a supercoiled DNA conformation that facilitates opening of the double helix.
The present study establishes the mitochondrial phage-type RNA polymerase encoded by the nuclear RpoTm gene to recognize mitochondrial promoters of diverse architecture in vitro. It thus provides the first direct linkage in the plant kingdom between an RNA polymerase activity initiating transcription at mitochondrial promoters and a nuclear gene encoding a mitochondrial phage-type RNA polymerase. In vitro transcription studies strongly suggest that at least for the diverse promoters shown here to stimulate initiation by RpoTm, promoter sequence specificity is conferred by the RpoTm core enzyme and does not require auxiliary factors. This questions previous ideas of different transcription factors being involved in transcription initiation at different promoter types controlling distinct genes or groups of genes (Mackenzie and McIntosh, 1999); compare also IV.1.2). Rather, general RNA polymerase cofactors and DNA-binding proteins may be involved in transcription in Arabidopsis mitochondria.
A comprehensive study directed at identifying mitochondrial DNA-binding proteins and transcription factors in Arabidopsis mitochondria will combine the computational prediction and preliminary biochemical characterization of mitochondrial DNA-binding proteins encoded by the Arabidopsis genome with a proteomics approach analyzing polypeptides attached to mitochondrial nucleoids (Thirkettle-Watts and Finnegan, 2005). Candidate transcription factors and DNA structure-modifying proteins emerging from this study could be used to complement the minimal Arabidopsis in vitro transcription system described here.
↓82 |
In vitro transcription studies provided little information on the role of RpoTmp in Arabidopsis mitochondria. Unlike RpoTm, RpoTmp did not display a marked preference for mitochondrial promoters as transcription initiation sites. However, non-specific transcription of DNA templates was far more efficient by RpoTmp than by RpoTm (Kühn, 2001); this study). Different RNA synthesis rates and promoter specificities may be related to the in vivo tasks of RpoTm and RpoTmp although possible effects of the thioredoxin tag on RNA polymerase activity, which might differ between the two recombinant enzymes, should be considered. The necessity to study in vitro RNA synthesis by untagged RpoTm and RpoTmp is discussed below (IV.4).
The data presented here on RpoTm- and RpoTmp-dependent RNA synthesis in vitro are in favour of different and complementing roles of the two RNA polymerases in transcriptional processes in Arabidopsis mitochondria. Promoter recognition by RpoTmp was confined to Patp6-1-916/913, Patp6-2-436 and Patp6-2-507, which nevertheless induced transcription initiation by RpoTm with higher efficiency. RNA synthesis from these promoters by RpoTmp was markedly exceeded by non-specific template transcription. Moreover, nine out of twelve promoters utilized by RpoTm did not activate RpoTmp in vitro (Table 12). In the light of the in vitro initiation at diverse mitochondrial promoters by RpoTm it is unlikely that the apparent lack of promoter specificity of RpoTmp is due to the absence of specificity factors in the in vitro transcription assay that would otherwise enable RpoTmp to efficiently recognize identical promoters. The weak preference displayed by RpoTmp for Patp6-1-916/913,
Patp6-2-436 and Patp6-2-507 over random initiation sites may be conferred by conserved structural elements of phage-type RNA polymerases.
Mitochondrial as well as plastidial phage-type RNA polymerases have been observed - though not in the present study - to initiate transcription at the T7 promoter in vitro(Kühn, 2001;Lerbs-Mache, 1993). The sequence of the T7 promoter is entirely different from those of promoters recognized by plant phage-type RNA polymerases in vivo. In turn, T7 RNA polymerase has been suggested to initiate transcription at plastidial promoters of phage-type RNA polymerases following the observation that genes controlled by this promoter type showed enhanced transcript levels in transgenic tobacco overexpressing a plastidial T7 RNA polymerase (Magee and Kavanagh, 2002). It thus appears that the ability to preferentially initiate transcription at certain specific or non-specific sequences is a property that is conserved among phage and phage-type RNA polymerases, regardless of whether or not these enzymes require auxiliary factors for transcription initiation in vivo. Yeast Rpo41 has been proposed to possess a polypeptide region corresponding in structure, though not in sequence, to the T7 RNA polymerase specificity loop involved in promoter binding (Matsunaga and Jaehning, 2004); compare Figure 5). Corresponding structures might be formed by plant RpoT polypeptides and could contribute to an intrinsic ability of RpoT enzymes to preferentially initiate transcription at particular sequences in vitro (see Annex C for T7-based structural models illustrating a hypothetical folding of the RpoTm and RpoTmp polypeptides). The specificity loop region of organellar phage-type RNA polymerases shows limited sequence similarity to the T7 enzyme but is flanked by two highly conserved sequence domains. It is rich in positively charged and hydrophobic residues, supporting this structure to be involved in protein-nucleic acid interactions (Figure 32). With the aim of defining structural elements of RpoTm that confer promoter specificity, site-directed mutagenesis of recombinant RpoTm and RpoTmp could be carried out, of which the RpoT sequence region corresponding to the T7 RNA polymerase specificity loop would be a particular target. Such mutagenesis studies might well complement phylogenetic analyses of plant RpoT enzymes. Besides, it would be of major interest to compare in vitro transcription activities of recombinant RpoTm and RpoTmp to those of enzymes that are active in mitochondrial lysates prepared from Arabidopsis in order to characterize functional differences between RpoT core and holoenzymes.
↓83 |
Figure 32: Comparison of the specificity loop region of phage and phage-type RNA polymerases. | ||
The upper aminoacid sequence alignment illustrates the level of sequence similarity between RNA polymerases of the T7 phage type in the specificity loop region (blue bar) and adjacent sequences. Shaded positions are conserved in 60 % (light grey), 80 % (grey) or 100 % (black) of aligned sequences. The lower alignment shows identical sequence regions, with amino acids shaded according to their physicochemical properties (hydrophobic, shaded light gray; positively charged, shaded blue; negatively charged, shaded red). Sequences were compared of RNA polymerases of the bacteriophages T7, T3, phiYeO3 and SP6, mitochondrial RNA polymerases from H. sapiens (h-mtPol), S. cerevisiae (sc-Rpo41), S. pombe (sp-Rpo41), and sequences of Arabidopsis (At-), Nicotiana sylvestris (Ns-), T. aestivum (Ta-) and Z. mays (Zm-) mitochondrial (RpoTm), mitochondrial/plastidial (RpoTmp) and plastidial phage-type RNA polymerases (RpoTp). The alignment was generated using the Multalin algorithm (Corpet, 1988) and refined manually. See Annex D for accession numbers. |
While a role of RpoTm in transcribing genes from mitochondrial promoters can be deduced from the in vitro transcription studies described here, the function RpoTmp in mitochondrial transcription in vivo remains unresolved. It is possible that in the presence of transcriptional cofactors, RpoTmp initiates transcription of mitochondrial genes in vivo, perhaps at promoters not recognized by RpoTm. It may be speculated that RpoTmp functions as a transcriptase priming mtDNA replication, and initiates transcription at as yet undefined sites associated with origins of replication. Alternatively, RpoTmp may have an as yet unreported transcriptional role that does not entail the recognition of identified mitochondrial promoters. Interestingly, preliminary data suggest both RpoTm and RpoTmp to transcribe RNA templates in vitro (data not shown). A similar activity has been reported for the T7-phage RNA polymerase (Konarska and Sharp, 1989;Konarska and Sharp, 1990), and has moreover been proposed to catalyze RNA replication in maize mitochondria (Formanova and Brown, 1997) and transcript 3’-end extension observed in both mitochondria and plastids (Zandueta-Criado and Bock, 2004); D. Gagliardi, IBMP CNRS, Strasbourg, France, personal communication), although the functional significance of the added sequences is unresolved.
Prior examinations of the roles of RpoTm and RpoTmp in Arabidopsis mitochondria, which predominantly addressed possible developmental or tissue-specific differences between RpoTm and RpoTmp functions, did not allow defining roles of the two RNA polymerases in mitochondrial transcription (Baba, et al., 2004;Emanuel, et al., 2005;Emanuel, et al., 2005). The recent characterization of a transgenic Arabidopsis line carrying a T-DNA insert in the last intron of the RpoTmp gene and apparently possessing no functional RpoTmp supported a role of the enzyme in early plastid but not in mitochondrial gene expression (Baba, et al., 2004). Data collected for both wild-type and mutant plants on plastid and mitochondrial mRNA levels at different developmental stages and under different light regimes were interpreted with difficulty, as the mutant displayed altered RpoTm and RpoTp transcript levels compared to the wild type. Phenotypic effects were therefore not directly attributable to the loss of RpoTmp function. The authors put the short-root phenotype of mutant plants down to not only the lack of RpoTmp but also reduced RpoTm mRNA accumulation in roots, but did not detect significant effects of the T-DNA insertion on the accumulation of mitochondrial transcripts. Based predominantly on the observation that in the mutant, the induction of several plastid genes in dark-grown mutant seedlings upon illumination was delayed, they proposed RpoTmp to be the key RNA polymerase transcribing organellar genes during early seedling development and favoured a role of both RpoTm and RpoTp at later developmental stages. However, a model in which RpoTm and RpoTmp act at different stages of development is not supported by a study showing essentially identical expression patterns of RpoTm and RpoTmp(Emanuel, et al., 2005). Baba et al. (2004) moreover suggested that RpoTm could substitute for RpoTmp in the mitochondrion, while RpoTp might partially compensate for the loss of RpoTmp in the plastid. In vitro transcription studies of RpoTm and RpoTmp are compatible with this idea.
↓84 |
In contrast to dicots, monocots possess only one mitochondrial phage-type RNA polymerase (see I.3.3.1). The available literature is bare of data hinting at a dicot-specific aspect of mitochondrial transcription that would necessitate an additional phage-type RNA polymerase in dicot mitochondria. Thus, it is not unlikely that functions of the monocot RpoTm are in dicots distributed between RpoTm and RpoTmp, although no data exist to substantiate this proposition. Studies of mitochondrial promoter utilization and transcription in transgenic Arabidopsis plants that are reduced in RpoTm or RpoTmp protein levels or entirely deficient in RpoTm or RpoTmp function could contribute to defining distinct roles of these enzymes in dicot mitochondria.
A recent investigation of the gene encoding the human mitochondrial RNA polymerase moreover directs the search for target genes of phage-type RNA polymerases to the nuclear genome. Remarkably, this gene has been discovered to additionally give rise to an N-terminally truncated polypeptide constituting a nuclear phage-type RNA polymerase (Kravchenko, et al., 2005). This novel enzyme appears to be responsible for the transcription of a large number of nuclear protein-coding genes and initiates transcription at sequences that resemble neither the human mitochondrial promoters LSP and HSP nor promoters recognized by the nuclear RNA polymerase II (Kravchenko, et al., 2005). One could speculate that the recruitment of phage-type RNA polymerases to transcribing organellar and nuclear genes may have occurred as one evolutionary process.
The in vitro transcription experiments described here strongly suggest the participation of auxiliary factors in mitochondrial RNA synthesis in Arabidopsis. Although the above described differences in template utilization between transcriptionally active plant mitochondrial extracts and RpoTm are reminiscent of the deviating template requirements observed between the yeast sc-mtTFB-Rpo41 heterodimer and isolated Rpo41 (Matsunaga and Jaehning, 2004), it is unclear whether a mtTFB homologue is present in the extracts which contributes to DNA melting or modulates transcription initiation in vitro and in vivo. The preliminary characterization of MetA, which apparently is the only mitochondrial mtTFB-like protein in Arabidopsis, through in vitro transcription experiments suggests that the role of MetA may not be equivalent to that of yeast and animalmtTFB (see IV.3.1). This is in line with phylogenetic analyses placing MetA and other putative mitochondrial mtTFB-like rRNA dimethylases of plants in one group with the yeast nuclear Dim1 methyltransferase but not with experimentally confirmed mitochondrial transcription factors such as fungal or mammalian mtTFBs (see Figure 17 and IV.2.2). The data presented here do however not unequivocally contradict MetA to act as transcriptional cofactor in vivo. It is possible that an interaction of recombinant MetA and RpoTm in the in vitro transcription assay, which according to analyses of yeast and animal mtTFB is critical for mtTFB function (Cliften, et al., 2000;Falkenberg, et al., 2002),was obstructed by the thioredoxin-hexahistidine tag attached to RpoTm N-terminus. Likewise, the N-terminal hexahistidine tag of MetA may have impeded MetA function, although similar tags attached to recombinant mammalian mtTFB were not reported to be disadvantageous (Gaspari, et al., 2004). Two-hybrid analyses examining a possible association of MetA with RpoTm or RpoTmp were not in favour of heterodimer formation (data not shown). Yet, fusion partners attached to MetA and RpoT proteins in these studies may have hindered interaction in the two-hybrid assay. On the other hand, a two-hybrid assay has been successfully used to monitor sc-mtTFB-Rpo41 interaction in yeast (Cliften, et al., 2000). In vitro transcription studies using untagged RpoT enzymes, which could be prepared from sources described in chapter III.3, will be required to more adequately examine the ability of MetA to interact with RpoTm or RpoTmp and to modulate RpoTm- or RpoTmp-dependent transcription. An Arabidopsis mutant line carrying a T-DNA insertion in the MetA coding sequence has recently become available from the GABI-Kat mutant collection (http://www.gabi-kat.de/). Mitochondrial transcript analyses of this mutant and of transgenic plants expressing interfering RNAs directed at MetA transcript levels may help elucidate a possible role of MetA in mitochondrial transcription. Alternatively, MetA may merely be functioning as rRNA dimethylase in Arabidopsis mitochondria, whereas cofactors mediating transcription initiation in plant mitochondria might not be of the rRNA dimethylase type. Presumed 18S rRNA methylation by MetA is presently investigated by
S. Okada and A. Brennicke (Universität Ulm). The possibility that in addition to MetA, a mtTFB homologue functioning as transcription factor exists in Arabidopsis mitochondria which escaped the BLAST searches conducted here, has been discussed in section IV.2.2.
↓85 |
How could RpoTm initiate transcription at promoters in Arabidopsis mitochondria? In the presence of the native mitochondrial nucleoid, which is composed of not only DNA but also DNA-binding proteins of as yet unknown identity, the DNA may have a conformation that allows the RNA polymerase to initiate transcription at mitochondrial promoters without the aid of additional auxiliary factors (Figure 33, model A). However, in vitro studies using plant mitochondrial extracts as a source of transcription activity imply that (a) soluble cofactor(s) may be involved in promoter-specific transcription initiation (Binder, et al., 1995;Fey, et al., 1999;Tracy and Stern, 1995). As in yeast mitochondria, these cofactors which may be functionally equivalent but not similar in their structure to yeast mtTFB, could associate with RpoTm prior to promoter binding (Figure 33, model B). The cofactor-independent in vitro specificity of RpoTm for diverse promoter sequences (see Table 12) suggests that in order to initiate transcription in Arabidopsis mitochondria, RpoTm associates with general auxiliary factors rather than promoter sequence-specific cofactors. Similar factors might enable RpoTmp to initiate transcription at mitochondrial promoters. Considering that both plant mitochondrial extracts and recombinant RpoTm recognize promoters of different architecture but not all in vivo promoter sequences in vitro(Binder, et al., 1995); this study), it may be argued that different promoters could require distinct modes of transcription initiation. Such mechanisms are however unlikely to represent a major means of mitochondrial gene regulation since studies of mitochondrial promoter utilization in vivo suggest a non-stringent control of transcription initiation(see IV.1). Thus, mitochondrial genes in Arabidopsis may not be individually regulated at the transcriptional level.
© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme. | ||
DiML DTD Version 4.0 | Zertifizierter Dokumentenserver der Humboldt-Universität zu Berlin | HTML generated: 06.09.2006 |